Holding Output and Safety Through an EBA Negotiation

In 2017 I sat through a wage round on a 300-strong site while the furnace kept running at 1300°C the entire time, because it had no other option. Nobody at the table could pause it to make a point. That single fact shaped every decision the plant made for the next eleven weeks, more than any clause in the draft agreement.

Most people outside the industry assume an EBA negotiation is a industrial relations problem with a production side-effect. On a container glass site it's the other way round. The negotiation is the side-effect. The furnace is the constraint, and it does not care what HR and the union are discussing down the corridor.

Why the furnace decides your negotiating posture, not the lawyers

A melter running hot-idle at roughly 1300°C cannot drop below that threshold without risking glass freezing in the throat and forehearths. Trigger that and you're looking at a cold repair costing an estimated USD 10-20 million and three to six months of lost production. That number sits in the back of every plant manager's head the moment a bargaining unit starts talking about industrial action, and it should sit in the back of the negotiating team's head too.

Discrete manufacturing can pause a line, protect the asset, and resume when the dispute settles. Container glass can't. That's the fact OEM-affiliated consultancies routinely miss when they get pulled into a negotiation-period review, because they're trained to see a throughput problem and prescribe an equipment fix. It isn't a throughput problem. It's a campaign-life problem wearing a throughput costume.

Skeleton-crew hot-idle operation still needs minimum batch-house, furnace and forehearth roles staffed on every single shift. Drop below that and refractory life on what's meant to be a ten to twelve year campaign gets shortened, permanently, whether or not anyone signs a strike notice. You can't get those months back with a better opening offer.

What actually degrades first when experienced hands come off the floor

The first casualty in a negotiation-period slowdown isn't the wage bill. It's crew experience density on the hot end. When your best IS-machine operators get pulled into caucus meetings, or when cover staff fill gaps left by reduced rostering, defect rates move fast and they move in a predictable order.

Gob weight control, normally held to about ±0.5g per parison, starts drifting under less experienced hands. That drift shows up downstream as checks and overpress defects before anyone's even looked at a quality report. Bird-swing faults creep in on gob delivery timing. Stones and cord don't spike from the batch house changing anything, they spike because the person who used to catch a marginal gob visually and correct on the fly isn't standing there anymore.

Plant managers I've audited report reject rates climbing one to two percentage points in the first week of reduced-experience crewing, against a normal hot-end target of 1.5-3% of gob count. Effective bottles-per-minute on the IS machine, typically running 300-600 bpm depending on container size, commonly drops five to ten percent in the same window. That's an early warning KPI, not a lagging one. If you're only checking it at month-end you've already lost the argument.

The negotiation doesn't damage your furnace. Losing your best operator off the floor for six weeks does. Those are two different problems and most IR teams only see one of them.

The mould change is where supervision leaks out first

Here's a number that gets missed constantly. A well-drilled crew changes a job or mould set in under thirty minutes. That window commonly doubles the moment supervisors get drawn off the floor into negotiation sessions instead of coaching the change. And it's not the mechanical steps that slow down, it's the judgement calls. Where's the mould preheat curve sitting against the 480°C ±10°C target. Is that a baffle mark from alignment drift, or is it a swab burn from a fitter running behind. Cover staff can follow a checklist. They can't yet make the calls that keep a changeover inside thirty minutes, and that gap is exactly what a systemised Job Change Tool is built to close, because it locks the recipe, the mould spec and the sequence so the call doesn't depend on who's standing at the machine that day.

I worked a plant in 2019 running Emhart 8-section machines on 1990s-era controls, mid-negotiation, where the night shift lost its most experienced hot-end superintendent to the bargaining committee three shifts running. Job changes that normally ran twenty-three minutes were landing at fifty-plus. Not a furnace problem. A handover problem. The 0600 handover misses night-shift swabbing data on most lines I see even in calm weeks, let's say seventy percent of the time, and under negotiation pressure that number gets worse, not better, because the people who'd normally chase it down are in a meeting room.

Annealing doesn't get a pass either

Under reduced throughput the temptation is to assume the lehr looks after itself. It doesn't. The annealing profile, typically running from about 580°C down to 50°C across the lehr, has to stay in spec regardless of pull rate, because under- or over-annealed ware fails residual-stress testing on the polariscope. For anything going into a carbonated or spirits fill line that's not a cosmetic defect. It's a pressure-container safety failure sitting on a supermarket shelf. OSHA's General Industry Standard 29 CFR 1910 on hot work and confined-space exposure around furnace and forehearth zones stays fully enforceable no matter how thin your manning gets, and it should. Nobody negotiates that away, and no plant manager should let a wage round create the illusion that they can.

What this looks like in the three regions we work

In the US, this dynamic has hardened over the past two contract cycles. O-I's Fit to Win programme idled furnaces including sites in Auburn, New York and Waco, Texas through 2023-2024, and Ardagh idled capacity at Winchester and Dunkirk in Indiana the same year. That drawdown landed right as United Steelworkers locals went into contract renewal, and strike-authorisation votes have followed at multiple O-I sites across the last three cycles. A plant walking into that kind of round with thin bench depth on the hot end is negotiating from a position it built for itself eighteen months earlier.

Europe has its own version, driven by energy rather than restructuring alone. The 2022-23 gas price shock, with EU spot prices running $8-12/MMBtu against subsidised Gulf feedstock historically under $2/MMBtu, forced O-I and Ardagh to idle furnace lines across Germany, France and the Netherlands. That compression fed straight into subsequent Tarifvertrag rounds negotiated with IG BCE, where inflation-adjustment clauses became the sticking point against double-digit German CPI. Add EU ETS Phase IV's Linear Reduction Factor climbing to 4.3% a year through 2027, and European plant managers are negotiating wages and defending furnace campaigns against a tightening emissions cap at the same time.

The Gulf plays out differently again. UAE Federal Decree-Law No. 33 of 2021 and Saudi Labour Law amendments restrict trade-union style collective bargaining, so EBA-equivalent friction in GCC container glass plants tends to surface as individual labour-court grievances rather than furnace-wide industrial action. That doesn't remove the operating risk, it just relocates it. A GCC operator I audited last year was managing the same crew-experience drawdown pattern as a US or European site mid-dispute, just without a single collective bargaining table to point at. The defect signatures on the hot end didn't know the difference.

What we tell plants before they sit down at the table

Most outside advisors split the negotiation and the plant-floor programme into two workstreams, one for HR and one for operations, and hand them to different teams. That's the mistake. The causal chain runs in one direction: crew experience drops, gob weight drifts, reject rate climbs, mould-change time doubles. You can't manage that chain from an IR playbook alone, and you can't manage it from an operations playbook that ignores what's happening at the bargaining table.

A management audit before the negotiation opens gives you a documented baseline on reject rate, changeover time and gob weight CV, so you know within a week if cover-staff crewing is costing you more than the wage claim itself. We've run this against 300-person sites before and the pattern repeats: the plant that loses the negotiation isn't the one that pays more, it's the one that let its refractory campaign and its quality data drift while everyone was looking at the other room.

Zaid Hassoneh built Lean Glass's approach to this off direct 300-person EBA experience, running the floor while the negotiating table sat two doors down. If your plant has a wage round coming and you want a baseline on where your hot-end variance actually sits before you walk in, that's worth a conversation before the first session, not after the ninth.